Google has introduced two new AI models designed specifically for robotics, building on its Gemini 2.0 technology. The launch aims to support the rapidly advancing robotics industry, which is increasingly benefiting from AI improvements.
The first model, Gemini Robotics, enables robots to generate physical actions as outputs, while the second, Gemini Robotics-ER, enhances spatial awareness and reasoning abilities for developers.
The move follows a significant AI breakthrough by robotics startup Figure AI, which recently ended its collaboration with OpenAI.
Google has tested its Gemini Robotics model on its bi-arm robotics platform, ALOHA 2, and believes the technology can be adapted for complex applications, such as Apptronik's Apollo robot.
Investment in robotics is accelerating, with Apptronik securing$350 million in funding last month, including backing from Google.
Google's AI models are designed for various types of robots, from humanoid machines to industrial units used in factories and warehouses.
Industry experts believe AI-focused robotics models will help startups reduce costs and bring products to market faster. Google has a long history in robotics, having acquired Boston Dynamics in 2013 before selling it to SoftBank four years later.